Building a novel machine learning solution.

On August 27th, 2019, our application for an ERDF grant was approved by the Investitionsbank des Landes Brandenburg, as part of their GrĂ¼ndung innovativ programme.

The goal of the project is to build a machine learning solution to help us enrich data about companies. The expected result is a fully trained set of neural networks, capable of doing a multitude of tasks: company categorisation, news extraction, and company scoring.

Company categorisation

The first part of the project is a tool which can take as input information about a company (such as its website, or a description), and automatically assign it one or more of our 800+ categories.

News extraction

The second part of the project is a tool which takes as input a news item, and which can extract from its body information such as which company the news is about, whether the companies mentioned in the article have entered some kind of new relationship (e.g. a partnership, or an acquisition).

Company scoring

The last part of the project is a tool which is able to confer a scoring to a company, given all the data we have collected about that company. The scoring is intended to help our users better evaluate companies when interacting with the database.